Prediction of team league’s rankings in volleyball by artificial neural network method

Abstract

In this study, an Artificial Neural Network (ANN) model that can predict future team rankings in male volleyball professional league was developed. Data used to develop the ANN model were obtained from 2013 to 2015 league tables. Wins, defeats, home wins, and away wins were used as input parameters and team rankings as an output parameter. There are only a few studies about predicting the match results using the ANN method. Related studies are mostly based on football, basketball and bowl games. There are no studies researching the prediction of team league’s rankings in volleyball. This is the first study about team ranking prediction in volleyball by the means of the ANN. The results showed that the most optimal ANN method was a single hidden layer 4-neurone model which had “logsig” transfer function, “trainlm” training function, and “learngmd” adaptive learning function. The accuracy rate of the most optimal model was 98%, meaning that team standing in a league table can be forecasted accurately using this ANN model.